NASA Technical Reports Server (NTRS) 20160004964: Flight Analysis of an Autonomously Navigated Experimental Lander for High Altitude Recovery

3.9 MB·English

by  NASA Technical Reports Server (NTRS)

Flight Analysis of an Autonomously Navigated Experimental
Lander for High Altitude Recovery

Jeffrey Chin1, Justin Niehaus2, Debra Goodenow3
NASA Glenn Research Center, Cleveland, OH 44070

Storm Dunker4, David Montague5
Airborne Systems, Santa Ana, CA, 92704
First steps have been taken to qualify a family of parafoil systems capable of increasing the survivability and
reusability of high-altitude balloon payloads. The research is motivated by the common risk facing balloon
payloads where expensive flight hardware can often land in inaccessible areas that make them difficult or
impossible to recover. The Autonomously Navigated Experimental Lander (ANGEL) flight test introduced a
commercial Guided Parachute Aerial Delivery System (GPADS) to a previously untested environment at
108,000ft MSL to determine its high-altitude survivability and capabilities. Following release, ANGEL
descended under a drogue until approximately 25,000ft, at which point the drogue was jettisoned and the
main parachute was deployed, commencing navigation. Multiple data acquisition platforms were used to
characterize the return-to-point technology performance and help determine its suitability for returning
future scientific payloads ranging from 180 to 10,000lbs to safer and more convenient landing locations. This
report describes the test vehicle design, and summarizes the captured sensor data. Various post-flight
analyses are used to quantify the system’s performance, gondola load data, and serve as a reference point for
subsequent missions.
I. Introduction
P
lanetary science missions often send payloads to near space using balloon gondolas for a fraction of the cost
of rockets. These balloon gondolas, however, are usually not reusable after their mission. Expensive flight
hardware frequently impacts the ground with high velocities and can sometimes land in inaccessible areas that make
them difficult or impossible to recover. A substantial cost and effort savings can be realized by guiding the balloon
gondola or its’ experiments to a convenient specified location, away from hazards, after the mission is complete.

The flight test is collaboration between engineers at
NASA Glenn Research Center (GRC), the Columbia
Scientific Balloon Facility (CSBF), and Airborne Systems
(AS) as a first step towards qualifying a GPAD system for
subsequent scientific missions. The parafoil canopies and
Airborne Guidance Unit (AGU) have an expansive flight
history for commercial and military missions, typically
Parafoil
deployed from around 25,000ft via static-line cargo plane
drops. GPADS of this scale have never been tested, and
scientifically documented, from altitudes of 108,000ft,
where the atmosphere is significantly colder and less
dense. This flight seeks to characterize the performance of
AGU a GPAD system under these harsher environmental
conditions and rapid descent speeds. The on-board avionics
will monitor the condition of both the navigation system
Figure 1. GPADS traditionally deployed from and the payload it carries.
aircraft.

1 Aerospace Engineer, Propulsion Systems Analysis Branch, Mail Stop 5-11, AIAA Member
2 Aerospace Engineer, Combustion Physics and Reacting Processes Branch, Mail Stop 77-5, AIAA Member
3 Electrical Systems Engineer, Diagnostics and Electromagnetics Branch, Mail Stop 86-5
4 Parachute Systems Engineer, Airborne Systems, CA, US, AIAA Member
5 Senior Software Engineer, Airborne Systems, NJ, US

1
American Institute of Aeronautics and Astronautics
The ANGEL system consists of the
drogue, parafoil, AGU, and the Balloon
Flight System payload (BFS) subsystems,
where the first three form the GPAD
system. The individual components and
systems overview, as packed, are
illustrated in Figure 2.

This document describes the post-
flight data analysis of the information
gathered by the ANGEL avionics
platform. This data is used to validate
engineering analyses, document lessons
learned, and act as reference data for
future flights where GPAD systems could
provide appreciable benefits. The stored
data is used to verify thermal and
structural models, as well as validate
electrical power system and telemetry
performance. The information also serves
to measure the gondola’s descent stability
and characterize the parafoil performance.
The on-board data acquisition platform
Figure 2. ANGEL system overview.
recorded environmental conditions,
measuring the avionics performance and
verifying expected impact forces.
II. Vehicle and Mission Description

Figure 3. Balloon launch.
2
American Institute of Aeronautics and Astronautics
The ANGEL system was launched as a piggyback payload on a CSBF Mission of Opportunity on September 4th,
2015 from Fort Sumner, New Mexico. The 6,000lb max gross-weight host vehicle, nicknamed “Thunderbird”, was
lifted by a 29 million cubic foot Helium balloon envelope pictured in Figure 3. Figure 4 shows Thunderbird in white
with the ANGEL system mounted front and center. The Airborne System drogue, parafoil, and AGU are the yellow
components, sitting on top of the Balloon Flight System (BFS) painted orange.

The BFS gondola, designed and built at
NASA Glenn Research Center, is
composed of a beveled rectangular
aluminum skeleton reaching a total system
weight of 230lbs including the AGU and
canopies. During integration, the drogue,
parafoil, and AGU are stacked vertically on
top of a honeycomb spacer and connected
via four tie-down D-rings. The BFS is
attached to the Thunderbird gondola by a
single eyebolt on the top surface and
restricted from motion after release using a
guide rail on the side of the gondola seen in
Figure 5. An interference pin mounted to
the gondola engages the BFS eyebolt. In
order to release ANGEL, a linear actuator
pulls the interference pin, disengaging the
eyebolt, see Figure 6, allowing the ANGEL
system to fall from the host vehicle after
reaching its floating altitude. After release
of the BFS, a four-roller guide on the side
of the BFS, shown in red in Figure 5,
travels along the rail on the gondola,
minimizing pitch, roll, or yaw motion, and
improving the likelihood of a straight
Figure 4. Thunderbird Gondola.
trajectory for ANGEL during release. The
drogue static line is fastened through a hole in the top of the separation device housing. During release, the drogue
static line is pulled taught and the drogue bridle, drogue, and attenuation stitching are extracted. An attenuation strip
applies a near constant force to the apex of the drogue for the first approximately 70 ft of descent. This prevents the
drogue from recoiling into the BFS after release and allows the system to build airspeed. Approximately four
seconds of freefall was predicted as necessary to build sufficient dynamic pressure to inflate the drogue.

ANGEL then intended to descend under drogue for approximately 400 seconds, keeping the payload stable and
below Mach 1 until it reaches a GPS trigger altitude of 6,250m (20,505ft) above the programmed IP elevation. For
the area being tested, the trigger altitude is about 7,620m (25,000ft) MSL. At this point the AGU commands the
release of the drogue using a dedicated motor, and the drogue is jettisoned, while simultaneously deploying the main
parafoil. The drogue and deployment bag separate from the main canopy by design. The AGU then commences
GPS navigation using standard GPADS flight software. The parafoil is controlled using motors contained within the
AGU. In the event of a GPS lock loss, a timer backup was present to trigger the drogue release and main parachute
deployment.

All active control sensors were located within the Airborne Guidance Unit, with all redundant data acquisition
systems contained within the gondola. Redundant GPS and multiple cameras were placed on the corners of the top
plate, with streaming 2-way telemetry antennae located on the bottom face. Live two-way telemetry and video feed
were used to monitor and log the system status and trajectory. Data was gathered starting in the pre-launch segment,
through ascent and descent and logged to multiple on-board storage devices and ground stations.

3
American Institute of Aeronautics and Astronautics
Figure 5. (Left) Front side of the BFS with separation carriage in red.
(Right) Actual system being mounted to the host vehicle guide rail.

Figure 6. The payload is secured vertically from a single 1" diameter steel pin.

4
American Institute of Aeronautics and Astronautics
III. GPADS Objectives
Concerning the advancement of GPADS technology to space recovery applications, the ANGEL test has
provided the opportunity to test this GPAD system at conditions not previously possible. Specifically, the
challenges to surmount include:
 Operating electro-mechanical systems at combined low densities and low temperatures
 Textile tolerance to radiation exposure
 Safely and reliably releasing the payload and the subsequent inflation of the drogue
 Remaining under Mach 1 and transonic regimes during droguefall
 Using GPS sensors at or beyond their altitude limits
 Mission Planning for large offset balloon flights
A. Thermal Management

Due to the combination of low temperature and low density, design attention was given to the battery packs in
the AGU. Too low of an operating temperature will quickly drain and potentially damage the Li-ion batteries. A
common minimum safe operating temperature is -20C. The GPS receiver had a minimum operating temperature of -
40C. Accordingly, the airborne guidance unit included a thermostatically controlled battery heating system. This
configuration was designed and environmentally tested to function over the ascent and descent to/from 100,000-ft
MSL. The warming function would turn on when the internal temperature dropped below a pre-selected
temperature. The GPS receiver chip was not directly thermally controlled, but confidence was provided in the form
of successful environmental tests.

It was thus a critical objective that the batteries health to be maintained throughout the flight envelope. From
previous flights at lower altitudes, where the thermal management system was first tested, overheating due to
reduced thermal capacity of air was not expected to be an issue. The AGU draws relatively little power during
ascent, and during descent, wind chill plays a dominant role.

B. Textile Exposure

Textiles react strongly to UV light exposure.
Normal exposure rates at earths surface can require
days to observe degredation to material properties.
However, at higher altitudes, with less filtering
performed by the atmosphere, increased exposure to
light and radiation sources were expected to accelerate
the degredation process. As a result, lighter colors
were selected for textile components since they absorb
less light, material strengths were intentially overbuilt,
and a minimum of structural elements were left
uncovered by deployiment bags. It was thus
anticipated that significant tolerance to material
degredation was present in the design of the system.
However, longer duration missions will likely require
light and radiation barrier protection until the time
when the system would be used.

A GPADS objective was therefore to demonstrate
that the materials selected and methods of textile
protection were adequate for short duration balloon
missions.

Figure 7. ANGEL Packed and Rigged
5
American Institute of Aeronautics and Astronautics
C. Safe Release

Of critical importance to the ANGEL flight was to have a safe and controlled release from the host vehicle.
Fouling a release could result in a loss of all mission objectives. On release of the ANGEL mass, in addition to
some balloon dynamics, the gondola rotates so that the new CG (less the ANGEL system) is directly beneath the
balloon tether. Initially conceived separation mechanisms posed a potential risk for the lowest portion of the gondola
to possibly re-contact the BFS or parachutes by rotating into them as they fell by. With this ‘kick’, the ANGEL
system could theoretically rotate as a mass in motion for 4 or more seconds until a suitable stabilizing force was
available from the drogue, by which time, the drogue bridle could be wrapped up and tangled on the payload.
Furthermore, due to the lack of real estate on the top face of the BFS and the requirement to keep the parachutes
clear of obstruction, all vertical forces were concentrated through single cantilevered eyebolt and pin. A dual pin
system (one in each back corner) was considered, but ultimately deemed more likely to fail or induce rotation upon
release. The final release design implemented the single pin and rail system, previously discussed, minimizing any
rotation induced by the abrupt release.

Parachute deployments in low density atmosphere are known to have unique inflation characteristics.
Particularly important is the need to manage rebound after elongation of the flexible elastic textile structure (risers,
lines, and gores) due to conservation of momentum between payload and the drogue itself. There are documented
cases where the parachute has rebounded to be below the payload (i.e. closer to earth than the payload), which is
especially dangerous as it could tangle on the payload itself and compromise the entire flight. The design of
deployment was therefore important. As a preventative measure, a load attenuation strip was added between the
drogue and the drogue deployment bag. This attenuation strip would peel for about 85 ft of payload descent. A
longer distance was desired, but not permitted at the time. It would be desirable to maintain attenuation until
sufficient dynamic pressure could be achieved to inflate the drogue.

A GPADS objective for the ANGEL drop was to prove the pin and rail release systems by having a clean release
from the host gondola, with no rotation motion or re-contact with the gondola after release. Further, it was also an
objective that the drogue remain above the payload at all times.

D. Descent Speed Control

Fundamentally, a decision was made early to avoid Mach 1 and
transonic speeds to prevent sonic boom shock wave, forebody vibrations,
changes to drogue performance, and other challenges. At the time of
architecture selection, the planned release was to be from 130,000ft. A 9.85
ft D ribbon parachute was selected as the preferred drogue which would
O
ensure velocity < Mach 0.75 if dropped from 130,000ft. The release altitude
was later reduced to 105,000ft, which only provides more margin. This
drogue is a mature qualified design which has a known performance at
lower altitudes. A secondary objective of the ANGEL drop was to
characterize the drogue’s opening, stability, and descent rate performance at
these yet-tested higher altitudes.

To understand the droguefall performance with an ANGEL
representative payload configuration, a C characterisation test was
D
performed. This information was important to ensure both Mach limits
were maintained as well as calculating the correct droguefall time for the
backup timer functionality.
Figure 8. Drogue CD Characterization

Test
E. GPS in Adverse Conditions

Because the planned altitude of the test, there was uncertaintly how the U-blox GPS module would perform. As
a result of the concern wether the AGU would lose GPS lock at higher altitudes and regain it, or not, during high-
speed descent, a backup timer was added as an either / or condition with the 25,000ft MSL GPS altitude trigger. An
6
American Institute of Aeronautics and Astronautics
objective of the test was then to prove out the GPS functionality and the programming logic for deployment of the
main parachute. GPS reception was also a concern given the presence of multiple other communication systems and
the large metal body of the vehicle itself. During integration it was discovered that an external antenna and filter was
necessary to maintain GPS lock when exposed to CSBF communication signals.
F. Mission Planning

Mission planning is routine for existing commercial applications of GPADS. However, special efforts were
required to assist the ANGEL mission. Due to the nature of balloon flight, where the system blows with the wind, it
is never precisely certain what the ascent trajectory will look like, or where the release location will be.
Consequently, numerous hazard restrictions were levied onto ANGEL. Hazard avoidance included the following
restrictions, shown in Figure 14:

 8 nm diameter No-fly zones near population centers – entire failure footprint (red circles)
 4 nm keep out zones from roads, highways, interstates, and high tension power lines – entire failure
footprint
 1 nm diameter keep out zones from dwellings (inhabited or uninhabited) – ballistic failure footprint (purple
circles)

As a result of the uncertain ascent path, there was the expectation that release would be decided on-the-fly when
the ballistic footprint was acceptable. Therefore, the decision to release places a higher priority on safety than
accuracy of landing.

Prior to the first launch window, suitable impact points (IP’s) were sought and surveyed, based on anticipated
ascent trajectory. This involved a day or two of scouting safe, accessible land with landowner consent. However
for the actual launch day there was not sufficient time based on changed wind conditions and personnel availablility
to scout for IPs.

The GPADS objective for mission planning was for the ANGEL to land within 150 meters of a pre-selected
landing point based on the predicted ascent trajectory and release point, while successfully avoiding the hazards
identified.

7
American Institute of Aeronautics and Astronautics
IV. Passive BFS Data Acquisition System
Balloon Flight System Components:
1. NASA Gondola
2. NASA Impact Sensors
3. NASA Battery
4. NASA PMAD
5. NASA Avionics
6. NASA Radio
7. NASA GPS
8. Airborne Systems Antenna
9. NASA Antennae
10. Thermal Foam
11. Go-Pro Camera Systems
12. Eye Bolt
13. Tie-Down Rings
14. Subsystem Shelf
15. Separation Carriage

Figure 9. (Above) CAD model showing the internal
arrangement of BFS sensors.
Figure 10. (Left) Isometric view, with insulation hidden.
Arrays of sensors were contained within the aluminum BFS
frame to record the flight from the perspective of the test
payload. This passive data acquisition system was used to
measure the GPAD performance and quantify the flight loads.
The platform consisted of a power source, radio, sensors, flight
computer, insulation, and ballast. A three-cell battery pack each
comprised of 5 series connected SAFT B0562 Lithium Sulfur
Dioxide batteries were enclosed within a passive thermally
controlled foam shell. Power was distributed from the battery
using a custom power management and distribution (PMAD) board connected to the data acquisition units, and radio
transmitter. The data acquisition units were comprised of the following sensors types:
1) GPS- Measure position, altitude, velocity, heading and time
2) Accelerometers- 3D accelerations, vibration loads and orientation
3) Gyros- Orientation and angular velocity
4) Magnetometers- Aid in calculating heading and IMU sensor fusion
5) Thermocouples- Monitor temperature of electronics
6) Cameras- Visual feedback
7) Current/Voltage Monitoring- Battery performance
Each of these sensors was redundantly measured from multiple locations and multiple power sources. All BFS
sensors were independent from the sensors inside the AGU that were used to actively navigate the payload on
descent. This system also periodically broadcasted and received telemetry data from a ground station located along
the projected mission path. The two secondary data acquisition (DAQ) systems passively monitored gondola
conditions from inside the blue structure in Figure . These sensors provided redundancy, along with a secondary
perspective on the flight loads while the AGU was suspended above the gondola during its final glide segment. One
of the internal DAQ systems also continuously streamed data to ground stations to help monitor the payload status.
This secondary radio system helped ensure maximum data recovery, even if the payload were to be destroyed on
impact.
8
American Institute of Aeronautics and Astronautics
V. Results
A. Day of Flight Conditions

A regional weather forecast was used to plan for acceptability of landing locations. The temperature, wind
speed, and wind heading conditions predicted for the flight are shown in Figure 11 below. The inherent canopy
forward airspeed of the parafoil, based on the suspended weight, is also displayed to demonstrate that the forecast
wind is about half of the speed of the system for the portion of altitudes using the parafoil.

Figure 11. CSBF Forecast for the Area.

9
American Institute of Aeronautics and Astronautics
B. Launch

After a two-hour prelaunch inflation period, the host vehicle was launched at 6:50am. Given the size of the
balloon and payload, pendulum effect during release was eliminated by driving the payload until it was directly
beneath the balloon at release. This ground vehicle, nicknamed “Big Bill”, can be seen in Figure 12 as the balloon
began to ascend.

Figure 12. View from just after Launch from Thunderbird.
During this release, the payload experienced a 3 g vertical acceleration and 3 Hz dampening vibration as shown
in Figure . At this point the internal temperature was approximately 90°F with a max rotational speeds of 0.26 rad/s
occurring.

Figure 13. Launch Vibrations.
10
American Institute of Aeronautics and Astronautics